OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

cgmo-webcgm message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: re: [cgmo-webcgm] QUESTION. Transparency (element) and alpha channel.


Lofton,

  > Attention color wonks....

  > At 01:01 AM 10/11/2005 -0700, Cruikshank, David W wrote:
  > >
  > >----------
  > >>T18.4  Transparency  I have a note that indicates there probably should 
  > >>be a note under Other indicating that this element should be prohibited 
  > >>when the colour model has an alpha channel?
  > >
  > >Why?  This element says whether (for example) gaps in dashed lines show 
  > >what is behind, or show the Auxiliary Colour.  It seems unrelated to the 
  > >presence of an alpha channel.
  > >
  > >This was a note I found in the document Dieter put together on WebCGM 2.0 
  > >for the Houston meeting.  Don't know the rational.

  > Recommendation.  Drop it.  TRANSPARENCY (element) and alpha channel are 
  > pretty much orthogonal concepts.

  > Question.  However, is there an issue about these following and alpha
  >  channel?

  > 1.) T.18.16 TRANSPARENT CELL COLOUR (v3/v4 element)

  > 2.) T.21.1  ESCAPE 22, Transparent Cell Colour (for v1/v2 metafiles)

  > 3.) T.21.1  ESCAPE 45, Alpha Transparency (modal attribute).

  > Some clarification might be in order.  To me:

  > a.) for #1 and #2, this makes sense to me:  if TCC is "transparent", then 
  > that rules over any alpha values on individual colors in cells; if 
  > "opaque", then the alpha values pertain.

  > b.) for #3 ... it seems to me that they should combine.  I.e., if the alpha   > channel for a color value (e.g., of LINE COLOUR) was 0.5, and if Esc45 
  > current value was 0.5, then shouldn't the effective alpha value be 0.25?

  > If #b is true, then the same rule actually answers #a -- they combine (by 
  > multiplying), where TCC can only have the values 0.0 and 1.0.

  > Thoughts?

I don't see them as orthogonal concepts.  TRANSPARENCY operates at the 
primitive level whereas alpha operates below that at the pixel level.

If TRANSPARENCY is on before a dashed polyline then the gaps between line 
segments are void or not drawn and so the alpha has no effect on these voids.
The same is true for transparent color cell. If a cell is determined to be transparent,
it is also void and likewise the alpha has no effect on that cell, that is whatever
was drawn prior to the cell array shows thru.

When TRANSPARENCY is off, then the gaps are drawn using the AUX_COLOR
and the alpha is applied to that just like it is applied to the dashed segment 
of the line.

That said,k I can see where the effect of these elements could be misinterpreted
so clarification would be good.

Regards, 
Don.
  > -Lofton.


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]